Four review findings across the seam and both adapters.
registerAdapter now returns a handle carrying replace(providers): the
candidate route set is validated in full before anything moves, so a
route another adapter owns leaves the previous registration intact, and
the swap itself is one synchronous section with no observable gap. pi-ai
uses it instead of dispose-then-register — the old shape dropped every
route when the new set conflicted, and its facts cache could then equal
the registry's, so reverting to a working configuration never re-applied.
Its registration facts are also sorted by provider, so a settings
document that merely reorders keys no longer triggers a swap.
DeepSeek's per-request snapshot now carries the credential facts, and
resolveApiKey receives it instead of re-reading the raw config: a
settings generation the resolver rejects can no longer contribute its
literal key to a request the previous generation's endpoint serves.
pi-ai only defers to the SDK's provider-native discovery when a profile
names no credential at all; a configured apiKeyEnv that misses now fails
with MISSING_CREDENTIAL naming the route and the reference, instead of
handing pi-ai undefined and letting it authenticate with an unrelated
ambient key.
The eager boot-time credential probe is gone: it could run before the
credentials service mounted and reported every failure as a missing key.
The route stays registered and browsable; the first request gives the
accurate error, whose guidance now leads with the credential store and
mentions a literal apiKey last.
The shipped surfaces loaded the harness home's .env into the process
environment before cordis booted. credentials-local then saw every stored
key as an ambient launch override: describe reported source 'env' with
writable false, and set/unset rejected as shadowed — so a key the web page
or TUI stored was unrotatable and undeletable from the next run onward,
and the adapter kept using the value captured at launch.
The home's .env is now the credential provider's own store, read by that
provider alone and hot-reloaded by it. The genuine launch environment and
the invoking directory's .env (loaded by the bin) remain the read-only
ambient layer, so a plain composition without the provider still resolves
keys exactly as before.
Proven by a real restart in the loader composition: store a key through
the seam, dispose the tree, re-boot over the same harness home, and the
entry is still file-sourced and writable — rotating it lands on the very
next request.
- gate model selection on steering-placement image carriers from enqueue
until their steering/message event publishes; release the gate when an
admission ends idle without publication (both behaviorally asserted)
- reject session.updateQueue edits carrying non-text blocks at the RPC
boundary (queue edits cannot bypass image admission)
- extend the durable-directory walk past a first-created DSH_HOME to the
deepest pre-existing ancestor
- strip Windows-style separators from attachment display names on POSIX
- verify attachment reads with a header-only probe (digest already proves
the bytes decoded fully at admission); document the read path
- make SessionInputShell.addImages refusal observable and keep workspace
transfers/composer intake from leaking refused drafts
- own ONE recursive image walk (dsh-llm contentHasImage) across apiproxy,
pi-ai, compact-basic, and the DeepSeek text-only assertion
- drop the redundant canonical-base64 regex and the no-op role read
- move AttachmentId/AttachmentError out of types.ts (brand.ts/error.ts);
document why AttachmentError does not extend HarnessError
- document the hard attachments inject in both consumer READMEs
apps/cli/cordis.yml gains settings-local, credentials-local, and the bare
dormant llm-pi-ai row (manifest deps added for the resolver contract);
llm-deepseek drops its !!js apiKey inline for per-request credential
resolution. Both adapters tag apiKeyEnv role('credential-ref') so the
form mounts the credential control. The web e2e scaffold isolates a
harness home per run — an in-process boot must never touch the
developer's real ~/.dsh — and the new models-settings scenario pins the
whole loop through the shipped app: dormant directory as add vocabulary,
schema-driven editor apply landing in settings.yaml, the route
registering live (topology frame), and a write-only key landing in the
temp .env with the configured badge converging. A hermetic test-owned
reference name keeps a developer's real provider keys from flipping the
badge. schema-form joins the platform module table (seed + externals)
so client bundles share one instance.
- llm-deepseek: the uncatalogued resolveModel fallback declares text-only
modalities — the wire route is text-only regardless of catalog
membership, so "unknown" must not let the host persist-then-fail images.
- session.selectModel also consults the pending-inbox mirror: a queued
image prompt enters the log only when claimed, after a switch would land.
- attachment store: ensureDurableDirectory syncs every ancestor entry up to
a caller-vouched boundary regardless of what mkdir reports — a raced
"already existed" is not "already durable".
- One image walker (imageBlockIn/imageInEvent) now serves both attachment
authorization and the selection gate; referencedImage therefore also
authorizes references inside wrapped message content.
- InputHub: the scope disposer resolves the conversation service optionally
(teardown/HMR must reach quiescence), and a send failing after its scope
died releases the in-flight drafts instead of restoring them onto a
disposed shell.
- http-bridge destroys declared-oversize requests with connection: close
instead of draining a body the client can trickle indefinitely.
- LlmService validates AND detaches modality arrays identically on the
advisory and exact routes; READMEs record the fourth INVALID_MODEL_INFO
rejection reason.
- CLI provider docs (JSDoc, README pair, Agent Note pair) describe the
reuse behavior; llm-route.spec now parses the SHIPPED cordis.yml through
the production extraction, pinning the row coupling.
- image-display lane pins gallery/rail shape in inline snapshots and the
object-URL scheme this environment must take; stale host.schema comment
dropped.
ctx.llm gains 'llm/adapters-updated' — a payload-free registry notification
emitted at every topology commit point (adapter routes registering or
disposing, directory entries appearing or withdrawing) with contained
observers and INVARIANT rethrow — plus registerConfigurableProviders/
listConfigurableProviders, the directory of routes an adapter plugin can
activate through configuration. llm-deepseek declares deepseek-official
(whole llm-deepseek section as profile); llm-pi-ai declares the full
installed catalog under providers.<provider> even while dormant, so the
web settings surface can offer every provider before any route exists.
The invariant companion asserts the registry stays readable at each
notification.
The native adapter's route was named deepseek, colliding with pi-ai's
catalog provider of the same name, so the two DeepSeek paths could never
be mounted side by side. The web settings page needs both configurable at
once. Compositions, fixtures, goldens, scaffolding defaults, and docs all
move together (pre-release, no shim); TUI/session-query-spill/
missing-credential goldens re-recorded through their keyless refresh
modes because provider-name length shifts box padding and spill
truncation points.
New credentials data-structure page (type-equiv manifested), group README,
rewritten llm-deepseek/llm-pi-ai READMEs (dynamic configuration, dict
profiles, credential chain), capability-seams/service-role registration,
Agent Note (bilingual), demo compositions mounting settings-local +
credentials-local with no inline key plumbing, installSettingsSection
consumer helper on the settings seam (deduplicating both adapters' wiring),
jscpd symmetry markers for the provider twins, runtime-closure additions for
python/sdk-runtime, and doc-budget ceilings AGENTS.md 1750→1755 /
packages/README.md 850→865 for the structural one-line group rows.
llm-deepseek gains a Loader+Include composition spec proving external
settings.yaml/.env edits reach the very next request, and a real-API e2e
where only a credentials-local document holds the key. The headless example
pins the first-run missing-credential UX as a keyless stream-json snapshot
(new credentials.cordis.snapshot.yml scenario); runLoaderSmoke learns
expectedExitCode so a designed failure surface can be pinned instead of
masked.
The adapter takes an options thunk and a per-stream credential resolver
instead of frozen construction facts: base URL, catalog, defaults, idle
budget, and the API key re-resolve at each operation, so a settings or
credential change reaches the very next request while in-flight streams
keep the facts they started with. resolveAdapterOptions is the one explicit
resolve step (entry config fails loud at load; a live snapshot failing a
beyond-schema bound keeps the last good options). The plugin layers its
entry config under the optional llm-deepseek settings section and resolves
keys literal-first through ctx.credentials with an ambient env fallback; a
missing key now registers the route, warns, and fails each request with
actionable MISSING_CREDENTIAL instead of failing plugin load. The
registration-captured retry policy re-registers the route in place when it
changes.
Implements the approved simplification Agent Note: sse.ts now pipes the
response body through TextDecoderStream and EventSourceParserStream
(eventsource-parser/stream) and keeps only the DeepSeek protocol shim —
yield each event's data, terminate on [DONE], throw
LlmError('STREAM_CLOSED') on EOF without the sentinel. The SSE
spec-conformance tests are deleted; sse.spec.ts pins only the
[DONE]/STREAM_CLOSED/EOF contract, including the new spec-strict verdict
that an unterminated trailing event is truncation (the old parser
flushed it — a robustness nicety no real provider shape needs).
eventsource-parser@^3.1.0 becomes llm-deepseek's second runtime
dependency (already in the lockfile transitively via the MCP SDK).
Docs: the Agent Note moves proposed/ → implemented/ and is rewritten per
the lifecycle contract; the rejected NIH roll-up note's inbound links
follow. The twin-adapters note, dsh-llm LlmAdapter JSDoc (and its
type-equiv fences), cookbook, group/package READMEs, root AGENTS.md
layout line, sdk-helper comments, and the regenerated config catalog
drop the "hand-rolled fetch + SSE" claim in both languages; all eight
touched pairs re-recorded.
A well-formed provider stream that ends with finish_reason stop and zero
content blocks previously became a successful empty assistant message: the
turn completed silently, and drivers like goal-session counted the no-op
round. Both adapters now map that degenerate completion to a finish
{kind:'error'} with the new canonical EMPTY_RESPONSE code from dsh-llm, and
dsh-llm-retry adds the code to its default retryable set, so the existing
closed-step recovery path retries it and fails loud once the budget is
exhausted.
Covered by adapter unit tests, an llm-retry default-policy test, and a new
authored keyless ACP snapshot (empty-response-retry) with a deterministic
1 ms zero-jitter retry overlay.
Yichen Jiang